resolve-issues
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEDYNAMIC_CONTEXT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_CONTEXT_INJECTION]: The skill uses the dynamic context syntax (
!command) inSKILL.mdto execute shell commands at load time. These commands (git status, git branch, git worktree, gh issue list, and gh auth status) are used to gather environment state and are diagnostic in nature.\n- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes untrusted data from external sources.\n - Ingestion points: The skill retrieves open issues using the GitHub CLI and reads repository-specific guidelines and templates from the file system (e.g., CONTRIBUTING.md, issue/PR templates).\n
- Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions embedded within the issue descriptions or template files.\n
- Capability inventory: The skill has the capability to manage git worktrees, create branches, perform commits, push to remote repositories, and trigger the create-pr skill.\n
- Sanitization: No input validation or sanitization of the retrieved issue content or repository files is performed before they are processed by the agent.
Audit Metadata